New Assessment Method of Alveolar Bone Grafting Using Automatic Registration and AI-based Segmentation

Abstract

Objective: The aim of this Technical note is to present a new assessment method of alveolar bone grafts after secondary alveolar bone grafting using automatic registration and artificial intelligence (AI)-based segmentation. Methods: A total of 7 Japanese patients (4 men and 3 women) with a unilateral cleft lip, alveolus, and/or palate, who underwent secondary alveolar bone grafting between March 2021 and August 2022 were evaluated. Computed tomography (CT) analyses were performed about 1 month before the surgery, and 1 day and 6 months after the surgery. All CT images were imported into a 3-dimensional analysis workstation. CT images from after the surgery were superimposed onto CT images from before the surgery, by automatic rigid image registration. The segmentation of bone tissues was automatically performed by the AI-based function. Grafted bone was extracted by subtraction of the bone tissue after the surgery from the bone tissue before the surgery. The volumes and Hounsfield units (HUs) of the grafted bones were calculated. The intraclass correlation coefficient (ICC) was reviewed to assess inter-rater reliability. Results: The ICCs (2,1) of the volumes and HUs measured by the observers immediately after the surgery were 0.95 and 0.99, respectively. On the other hand, the ICCs (2,1) of the volumes and HUs measured by the observers 6 months after the surgery were 0.81 and 0.57, respectively. Conclusions: Our new assessment method enables simple and quick evaluation of residual grafted bone after secondary alveolar bone grafting and demonstrated relatively high inter-rater reliability.